Myasthenia gravis (MG) is an autoimmune disease involving the thymus in which 85% of patients have antibodies to muscle acetylcholine receptors (AchR-Ab) that interfere with neuromuscular transmission and can cause severe, sometimes life-threatening, weakness. Thymectomy has been used world-wide to treat non-thymomatous MG, based on retrospective non-randomized studies. Prednisone (a corticosteroid) is also frequently used to treat MG. Both therapies are often employed together and both have adverse effects. Whether thymectomy benefits those who are also receiving prednisone is not known. To investigate the safety while comparing efficacy, we propose a multicenter, multiracial, international, single-blinded, 3-year duration clinical trial in which patients aged 18-60 years with generalized AchR-Ab positive non- thymomatous MG are randomized to receive extended transsternal thymectomy (ETTX) or no thymectomy. Both groups will receive prednisone administered by a 'blind' evaluator according to the same set protocol aimed at establishing the minimum dose needed to achieve and maintain Minimal Manifestation (MM) status. The primary endpoint will comprise response and toxicity using a composite measuring clinical course, short and long term toxicities based on: (a) clinical efficacy of therapy evaluated by the QMG weakness score (b) frequency of serious adverse events (c) the total dose of prednisone (Area under the Dose time Curve, AUDTC). A significant difference favoring ETTX would establish its clinical benefits in this patient population, and provide indirect evidence of the possible benefits of ETTX in patients not receiving prednisone medication. Conversely, failure to demonstrate a significant difference in the global or individual components of the primary endpoint would suggest that thymectomy is an unnecessary procedure in the population studied. Subgroup analysis may show whether benefits are confined to those who are prednisone naive at entry or in a particular age group. Thus the results will impact on current clinical practice. [unreadable] [unreadable]